Maximizing Edge Retention – What CATRA Reveals about the Optimum Edge

Thanks to Ed Schempp, Matus Kalisky, and Chin Lim for becoming Knife Steel Nerds Patreon supporters! Your support allows us to fund CATRA studies. 

CATRA

Update 1/6/2020: I have since written more articles about CATRA looking at the effect of steel type: Part 1 and Part 2

Cutlery and Allied Trades Research Association (CATRA) makes an edge retention tester that measures slicing of cardstock impregnated with 5% silica (sand). You can see a video of what the test looks like here:

The tester uses a fixed load, test speed, and stroke length. A typical test for a plain edge steel knife is 60 cuts with 50N load at 50 mm/s, which takes about 15 minutes [1]. The test is often used to compare different steels to determine the edge retention potential of each. However, the CATRA tester is not necessarily a device that compares steels; it compares knives. What I mean by that is that the edge angle, how the blade was sharpened, etc. all affect the measurement. Only if all of those variables are kept the same can steels be compared. Even then there is the possibility that with a different set of variables two steels may behave differently. In other words, a knife with a very thin finely polished edge may show Steel A is better than Steel B, but that may be reversed with a thick edge that has a coarse finish. We won’t know for sure until we do the test.

The Study Summarized in this Article

In 2012 a set of edge retention tests were performed by Wister Hill who commissioned the manufacture of a set of knives in 154CM and CPM-154. Though the tests were completed in 2012 the results have never been published. You can read an article about 154CM here: 154CM – Development, Properties, Use in Knives, and Legacy. These steels have the same composition but 154CM is produced by conventional ingot metallurgy and CPM-154 is produced by the powder metallurgy process. The use of powder metallurgy leads to a very different carbide structure, as the powder metallurgy version has much smaller and more evenly distributed carbides. The carbides are very hard particles that greatly contribute to wear resistance of steel. Based on their smaller size and better distribution we might expect the powder metallurgy version to do better in a thin edge. Here are micrographs taken from the steel used in the CATRA study:

Conventional 154CM

Powder Metallurgy CPM-154

Niagara Specialty Metals donated the steel, the blanks were then cut out and rough hollow ground, then heat treated by Peter’s Heat Treating.  Knifemaker Butch Harner then did the final grinding of the blades. The blades were sharpened by Jason Bosman using DMT diamond plates of 325, 600, 1200, or 8000 grit with a DMT angle guide. The knives were given different edge angles of 56, 50, 34, 27, and 20° total (half of that per side). The final angle of the edges was then tested with a CATRA geniometer. The knives used a flat profile designed to work easily with the CATRA tester:

Sharpened Edges

Here are two edge-on micrographs of the CPM-154 and 154CM knives. I believe these were sharpened to 8000 grit as the scratches are all under 3 micron. If I get confirmation of what the edges were resharpened to prior to the micrographs I will update this article. The edges are approximately 1.5 microns at the apex. however, there are “holes” in the 154CM edge which are about 4 microns wide which may be evidence of “carbide pullout” where a large carbide or string of carbides is lost during sharpening.

Ingot 154CM

Powder Metallurgy CPM-154

Effect of Edge Angle

With a fixed edge thickness the height of the final edge increases with lower angles:

Micrographs taken of the profile of the edges show that the target angles were achieved:

50°

34°

20°

Edge angle had by far the strongest effect on edge retention, much stronger than other effects such as PM vs ingot and the finish it was sharpened to. Here are graphs showing the performance of 20, 34, and 50° vs cut length for an individual cut and also for total card cut in mm:

You can see that the initial cut length with a smaller angle is considerably higher and that the difference holds basically to the end of the test. This finding is significant because some have speculated that lower angle edges start out sharper but a more obtuse edge lasts longer [2]. And with the high wear that occurs in the CATRA test it isn’t likely that the situation would reverse with even further cutting. The initial blunting rate is relatively rapid regardless of angle and it then begins to level out. The highest TCC measured was over 1000 mm with an angle of 20°, and this decreased all the way to under 100 mm with 56°. In a CATRA study by Bohler-Uddeholm [3] with a range of steels, but unspecified edge geometry or sharpening, 154CM was measured at 547 mm, and M390 was measured at 959 mm. The 547 mm value would be with an edge angle around 30° in this study if other parameters are similar. So if the edge angle of a 154CM knife is reduced from 30° to 20° then it can match or exceed a steel with 75% greater wear resistance.

There was a difference in final thickness of the apex of the edge after the knives were run through the CATRA test. The 20° edge was about 23 microns after the test, 34 degrees led to about 19-20 microns, and the 50° were around 16-17 microns. All of these images are of the ingot 154CM steel. So it appears that with a lower angle edge it can wear down to a larger apex and still maintain better cutting ability than a higher angle edge.

20° worn edge (23 microns)

34° worn edge (19 microns)

50° worn edge (17 microns)

 

Effect of Thickness Behind the Edge

Half of the knives were ground to a thickness of 0.010″ and half to 0.020″ prior to sharpening the final edge. Because the knives were ground by hand there was some variation in thickness both between the knives and along the edge of an individual knife. However, these variations were not recorded. Thinner edges, regardless of final edge angle, did have a small advantage in edge retention, as shown in the figure below. This is consistent with previous studies using a fixed angle but different edge thickness [4].

Effect of Sharpening Finish

The DMT grit that the knives were finished to does not show as strong an effect as was shown with edge angle:

The only test performed with 120 grit was with the 20° edge angle so that finish cannot be compared with the others. With only two tests performed with 1200 grit it is also hard to tell where it fits with the others. Otherwise it appears that 600 grit did marginally better than 320 or 8000 grit. In the past some have proposed two basic scenarios, either: 1) a coarser finish leads to superior slicing due to “micro-serrations” or increased surface area, or 2) finer finishes start out sharper and maintain that sharpness to cut overall longer. This study seems to indicate instead a peak finish of 600 grit, or approximately 25 micron finish. However, there are a couple potential complications to point out. In the blog Science of Sharp, micrographs are shown from different DMT finishes which demonstrates that the finer finishes did not lead to a smaller radius edge and instead the edges were more ragged with 8000 grit [5]. Perhaps that is what we are seeing in this particular study. The author defended his sharpening skill and proposed a mechanism for why diamond plates at high finishes do not sharpen as well as softer stones. No sharpness tests were performed on the untested blades of this study so it is not known whether higher sharpness was achieved with the higher finish. The images taken of the edges of this study were taken from different angles so there is no 1:1 comparison with the Science of Sharp blog, so it’s hard to tell if they have the same level of unevenness. 

600 grit DMT finish [5]

8000 grit DMT finish [5]

Conventional Ingot vs Powder Metallurgy Steel and Hardness

In an article published by representatives of CATRA, they reported suspiciously round numbers of 500 mm TCC for ingot M2 steel and 1000 mm for PM M2, both tested at the same hardness of 63 Rc; in other words, powder metallurgy led to double the edge retention [6]. That result was not confirmed in this study. The effect of powder metallurgy CPM-154 vs the conventional 154CM was much smaller than the effect of edge angle. Because of the superior response to heat treatment, the CPM-154 was marginally harder with an average hardness of 62 Rc vs the 61.2 Rc of 154CM. All of the CPM-154 was tested at 62 Rc while the 154CM tested between 61 and 62 Rc. Therefore separating the effect of processing vs hardness is difficult, but some analysis can be performed by focusing on the 62 Rc 154CM. The small difference in hardness between ingot and PM is probably sufficient to describe the small difference in edge retention observed between the PM and ingot steel, which was around 8-10%. The initial cut (mm) was approximately equal between the two steels, which makes sense with a similar finish and edge angle. The carbide pullout observed in the micrographs of the edges doesn’t seem to have affected the initial cut. Despite the very similar initial cut, the CPM-154 usually pulled away by a small amount with further cutting due to its higher hardness:

The harder PM steel was not better in every case though, the 8-10% improvement would not be possible to estimate without a sufficiently large dataset. The difference is small enough that it would be difficult to perceive without rigorous testing. You can also see that the 50° tests (~20mm initial cut) did not reveal a difference between 154CM and CPM-154. This may be simply because the total amount of cutting is so small that a difference is not perceptible, though another intriguing hypothesis would be that with a sufficiently large edge angle that the edge is better able to handle the larger carbides of the ingot steel. However, considering that the difference between them is more likely explained by the small hardness difference rather than the carbide shape/size this might be a reach. Indeed, if a plot is made comparing 62 Rc to 61-61.5 Rc it is almost identical to the PM vs ingot figure. There would probably be more separation due to hardness if we had a 2-3 Rc difference rather than 0.5-1.0Rc. 

Effect of Cryogenic Processing

Half of the knives were given a snap temper at 300°F for one hour prior to a cryo process at -300°F for 4 hours. A snap temper helps reduce the possibility of cracking during cryo but also stabilizes some of the austenite making cryo less effective. All were then given an upper temper at 960-1000°F, with the difference tempering temperature to maintain hardness regardless of the cryo process. The average hardness with cryo was 61.7 Rc and without cryo was 61.5 Rc, so the compensation by tempering was effective for maintaining similar hardness. The high temperature temper also helps convert retained austenite which makes cryo less necessary. Some have proposed that cryo only leads to an increase in wear resistance with lower tempering temperatures because of a theory involving eta carbide precipitation that would be lost with an upper temper. However, there are several fantastical studies out there claiming extreme improvements of wear resistance or tool lifetime with cryo processing (82.5 times the tool life!), seemingly regardless of how it was done [7]. So the comparison with and without cryo is interesting on its own:

Surprisingly the steel without cryo had slightly better edge retention, though the difference is small. The reason for the small difference is difficult to guess, perhaps it is just statistical noise. It may be that with removal of the snap temper, extended cryo time, or the use of low temperature tempering would lead to an improvement in edge retention from cryogenic processing. However, we can conclude that the use of cryo will not lead to an automatic improvement.

Conclusions and Further Testing

The figure above is the calculated “relative contribution” to a regression model in terms of impact on R², a statistical measure of how well a model fits experimental data. The relative contribution was calculated by Minitab. From this study it should be apparent that lower edge angles lead to superior cutting performance. PM vs ingot, cryo, and edge finish appear to be minor factors; Minitab analysis seems to put the contribution to edge retention solely on hardness rather than the PM/ingot difference. Hardness and edge thickness have less effect than edge angle but are still important. It may be interesting to compare a steel with lower carbide volume between its ingot and PM version to see if any difference is perceptible between those. It would also be a good study to look at another type of sharpening stones to see if it was indeed an effect of diamond plate sharpening that led to the peak behavior at 600 grit. I would also like to compare a wider range of edge thicknesses and hardness to get a better idea of those effects since they were over a narrow range in this study. More tests with different types of cryogenic processing would also be interesting. In future articles on edge retention I will compare CATRA testing with other reported edge retention tests, and also cover CATRA testing of different steels with (hopefully) consistent edge angle and sharpening finishes to see what steels are best for edge retention. 

---

[1] http://www.catra.org///pages/products/kniveslevel1/slt.htm

[2] Diéguez, José L., Javier Martínez, José E. Area, Alejandro Pereira, and José A. Pérez. “STATE OF THE ART IN THE PROCESS OF DEBONING AND SLICING OF MEAT IN THE FOOD INDUSTRY.”

[3] http://www.bucorp.com/media/CATRA_Test2.pdf

[4] Verhoeven, John D., Alfred H. Pendray, and Howard F. Clark. “Wear tests of steel knife blades.” Wear 265, no. 7-8 (2008): 1093-1099.

[5] https://scienceofsharp.wordpress.com/2015/03/01/the-diamond-plate-progression/

[6] Gregory, G., and R. Hamby. “Sharp edge cutting technology-a review of hand held and machine knives/blades and their sharp edge retention.” Surface engineering 16, no. 5 (2000): 373-378.

[7] http://www.pangas.ch/internet.lg.lg.che/de/images/Subzero_Treatment_of_Steels_en553_116014.pdf